| Author |
Message |
junction_crosstalk
Joined: 03 Dec 2002
Posts: 24
|
 19 May 2003 14:07 Combination of MEDICI and HFSS |
|
|
|
Hi,
Does anybody have project experience or example which needs to use device simulator (e.g. MEDICI) and electromagnetic solver (e.g. HFSS) (on the same project) ?
thanks,
JC
|
|
| Back to top |
|
 |
div
Joined: 17 Jun 2002
Posts: 73
|
| 19 May 2003 14:51 Reply1 |
|
|
|
junction:
I had put a similiar question before.
Please do a search if you need.
Basiclly, I want to compare the simulation of the same structure
using two different solvers: EM solver and Device solver.
But there is no direct response to my post.
If you are interested, please put more detail and
let's discuss this topic.
Regards
|
|
| Back to top |
|
|
Pim
Joined: 26 Jan 2002
Posts: 78
|
| 19 May 2003 18:07 Re: Combination of MEDICI and HFSS |
|
|
|
I too, would be very much interested in any comments on this topic. I intend to simulate floating-gate MOSFETs as well as dual-gate MOSFETS. Such components are not very often found in techlibs. Concerning the floating-gate transistors, the assessment of the electrical fields is an important question.
So I do not know what to go for to get the most realistic simulation results - a device simulator or an EM-simulator program?
/Pim
|
|
| Back to top |
|
|
junction_crosstalk
Joined: 03 Dec 2002
Posts: 24
|
| 19 May 2003 18:19 |
|
|
|
Hi Div,
I have seen your message in the other forum. My problem is as follow:
Just imagine that I have a Si substrate. On it there is an oxide layer. On the oxide there are 2 rectangle contact pad separated by a distance. Those 2 contact pads are connected to the Si substrate through contact openings. Let's say we apply a voltage(e.q. 1V) to one contact pad and measure the current occured on the other contact pad. By doing this we can find the parasitic resistance of the substrate.
Now, if we contact the contact pad from the backside through a via and a grid ball, surely we need electromagnetic field solver.
One thing I am doing is tho simulate both cases separately and combine them in circuit simulator. I don't know if this accurate enough.
JC
|
|
| Back to top |
|
|
junction_crosstalk
Joined: 03 Dec 2002
Posts: 24
|
| 19 May 2003 18:28 |
|
|
|
Please search IEEE paper by S.M.Sohel Imtiaz, Samir M. El-Ghazaly. The develop a software called CESS which is combination of device and EM simulator.
But I am not sure these paper are useful for you guys as they're for me.
JC.
|
|
| Back to top |
|
|
superluminal
Joined: 31 Jan 2002
Posts: 95
Helped: 1
Location: Inside a quantum well
|
| 20 May 2003 16:52 |
|
|
|
I think what you are asking for is a bit complex. The example you provided actually needs very wide & accurate knowledge of many of the parameters of ALL the layers tou want to simulate. This is the real situation which made a process simulator , device simulator & EM simulator.
To be more clear , if you have all information you need to describe (& solve) the EM fields behavior in the region + all information you need to describe ( & solve) the electron motion in these region you must have the solutions ( from both points of view ) so they converge to the same results ( say , for voltage of some point).
This is actually done , in device modeling by solving both Possion & Schrodenger equations in the region. This is not a simple task at all , but this become more demanding as device dimesions go lower & lower . You can observe this more clearly in heterostructure devices or in nanostructure device.
Actually , I feel I'm not clear enough !. The problem need more space to be covered & can not be done in a message. I recommend you to see this book " The Physics of Low-Dimensional Semiconductors " by J. H. Davis , Cambridge Uni. Press . It contains examples of how Poission & Schrodenger equations are coupled & solved for band diagram calculations. You can also check this site http://www.nd.edu/~gsnider/ & see the 1D Poission-Schrodenger solver that man made.It was very difficult & requires high sophisticated techniques - just for 1D ; solving numerically . Imagine how a 3D one will be.
|
|
| Back to top |
|
|
